The performance, and perhaps in at least some cases the viability, of some optical interconnect architectures can be sensitive to a temperature stability of the optical components. For example, a silicon photonics Wavelength Division Multiplexed (WDM) architecture, which shows promise to provide a high bandwidth and low power optical solution, may require a rather precise temperature control (e.g., on the order of within about 0.5° C. variation) for reliable performance. The requirements of a thermal management solution capable of providing such a temperature control are challenging not only because the power dissipated as heat in the optical components must be effectively controlled, but also because both ends of the optical path may need to be at or near the same temperature as the control temperature to help ensure the proper control. These requirements are further complicated by the start-up phase of the system, where all the components are at essentially room or ambient temperature and will then heat up to their operating temperature. Additionally, the temperature of the individual components may vary based upon utilization (idle to peak or thermal design power (TDP)) or even natural or programmed room temperature changes.